The Motorola 56000 series are fixed-point digital signal processors. The 56000 and 56001 processors have 24-bit data and instructions, and operate at a maximum clock rate of 40 MIPS. The 56100 processor has 16-bit data and instructions, operates at a maximum rate of 30 MIPS, and has analog/digital and digital/analog converters integrated on the chip. The 56301 has 24-bit data and instructions, operates at a maximum rate of 80 MIPS, and has several built-in input/output interfaces. Although the processors have pipelines of different lengths, the assembly code is backward compatible. The CG56 domain generates assembly code for the 56000 processor and has been tested on the Motorola simulator and on a 56001 board.

Since the 56000 processors are fixed point, the floating point data type has no meaning in the CG56 domain. Fixed-point values can take on the range [-1,1). The most positive value is

for the 56000 and 56300, and for the 56100. The domain defines a new constant ONE set to this maximum positive value. In this chapter, whenever data types are not mentioned, fixed-point is meant. The complex data type means a pair of fixed-point numbers. The complex data type is only partially supported in that it is not supported for stars that have anytype inputs or outputs, except for fork stars. Integers are the same length as the fixed-point representation. Matrix data types are not supported yet.

Some of the demos use the Motorola 56000 assembler and simulator. You do not need to have a 56000 chip to run the simulator demos, the assemberl and simulator are available for downloading from Motorola at http://www.mot.com/SPS/DSP/developers/clas.html.